Designing for Retention: Habits and Recurring Value Without Tricks
Designing for retention means people return because the product serves them. What the evidence says about habits, and why the 66-day figure is not a target.
Why is retention a strategy decision?
Because it measures whether the product has a place in someone’s life. A team can win many sign-ups with campaigns, but if people do not come back, the cost of acquiring each user never pays off. Retention connects design to the business directly, which is why it should be treated as a strategic question, not as a screen adjustment.
The underlying question is simple: how often does this person need the result the product offers? A personal finance app may be needed every week. Tax software, once a year. Designing for retention starts by accepting that difference, rather than imposing the same rhythm on everyone.
What does the evidence say about habits?
The most repeated number is 66 days, and it is worth knowing where it comes from. Lally and colleagues followed 96 university volunteers, mostly postgraduate students, who repeated a daily healthy behavior for 84 days. Only 82 had enough data, and 39 fit the model. The reported range was 18 to 254 days, with a median of 66, but that median corresponds to that subgroup (The Behavioral Scientist: How long does it take to form a habit?).
The study measured self-reported automaticity with a habit questionnaire, not habit in the strict sense: the authors themselves say they assessed the development of automaticity in performing the behavior. In a January 2026 interview, first author Pippa Lally answered that the 66 days do not apply to most people most of the time, and that how long habit formation takes is highly variable (University of Surrey: Does it really take 66 days to form a habit?).
Two lessons are useful for design. The first: missing a day does not destroy the process. According to the British Psychological Society’s summary, a single missed day had little impact on later automaticity gains. The second: too many missed repeats, even if spread out over time, had a cumulative effect and reduced the maximum automaticity level (BPS: How to form a habit). Overall regularity counts, not perfection. Neither lesson proves that an app should use notifications to bring people back.
What is recurring value?
It is the benefit a person gets each time they return. It is not the same as usage. An app that gets people to open the screen every day without solving anything has high usage and low value, and sooner or later people uninstall it. Recurring value shows up when the person completes a task that used to be hard, or gets an answer that helps them decide.
To find it, ask what result people get in their first session. If the answer is vague, the problem is with onboarding patterns, not retention: the person has not yet understood the value. Without that first success, no return strategy works.
How do you design the return?
These practices are starting points. None of the sources in this article measures their effect, so validate them with your own product:
- Show real progress. A summary of what the person accomplished since last time, with figures they can verify.
- Leave unfinished work in a clear place. If a task is halfway done, the screen should show where it stopped and how to continue.
- Offer value outside the app. A useful alert, such as a renewal reminder, that the person appreciates because it saves them a problem.
- Respect the natural frequency. If the need is monthly, communication should also be monthly.
Streaks and points can raise activity in the short term, but the risk is that people return out of fear of losing a streak, not because the app serves them. That resembles the patterns a team should avoid; review dark patterns before adding them.
When should you NOT measure retention as a goal?
- When the product is used once by design, such as a one-time procedure tool. There, a good product is one that is used rarely.
- When the team has not defined what “returning” means. Opening the app, completing a task, or paying again?
- When the number is compared with another industry. Product categories have very different usage frequencies, so a figure without context says little. Compare with your own history.
Illustrative example
Illustrative example: an expense receipts app has low 30-day retention. The team first debates adding daily notifications. Before that, it checks what happens in the first session: 70% of people upload a receipt, but almost nobody sees the monthly summary the app calculates. The change is not a notification; it is moving the summary to the home screen and showing how much was saved this month compared with last. People return at the end of the month to review their summary, because it serves them. Retention rises, and the team did not have to ask for space on the lock screen.
To decide what the team does with this metric, the conversation with the product manager should include which decision would change because of it. And product strategy and UX strategy define which value the team wants to sustain.
Common mistakes
- Treating 66 days as the habit target for the product.
- Measuring usage and calling it retention.
- Using streaks and guilt to force a return.
- Pushing the same frequency to every user.
- Ignoring the first success and designing only for the return.
- Comparing retention with numbers from another industry.
What you can do tomorrow
- Define in one sentence what it means for a person to “return” in your product.
- Ask what result people get in their first session, and measure how many achieve it.
- Look at one month of data and separate people who returned after completing a task from those who only opened the app.